Install any skill in seconds. Free to start, no credit card required.
Get Started Free →[COMMUNITY] Generate an Official Languages Act review — Parts IV (services), V (language of work), VI (federal language obligations); service-equivalence matrix EN/FR; bilingual public-facing surface; active offer; Translation Bureau pipeline; OQLF acknowledgement where federal-Quebec overlap applies.
.claude/skills/thomasmoreai-arckit-ca-ola/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 149% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 50% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 37% | 0% |
> ⚠️ Community-contributed command — not part of the officially-maintained ArcKit baseline. > Output should be reviewed by qualified Canadian counsel and the relevant departmental authority > (ATIP coordinator, ITSEC officer, OCHRO language lead, CIO branch) before reliance. Citations > may lag current text — verify against the Justice Laws Website and the issuing TBS / CSE / OPC source.
text$ARGUMENTS
You are an enterprise architect generating an Official Languages Act review for a federal digital service.
projects/000-global/ARC-000-PRIN-*.md (federal principles, if present)ca-fitaa artefact if present (any public register triggers Part IV bilingual obligations).arckit/templates/_partials/RENDERING.md.arckit/templates-custom/ca-ola-template.md (user override).arckit/templates-custom/ca-ola-template.md.arckit/templates/ca-ola-template.mdscripts/bash/generate-document-id.sh <PROJECT_ID> OLA --filename for the artefact filename.<!-- DOC-CONTROL-HEADER --> marker per RENDERING.md. Use the Canadian classification scheme (UNCLASSIFIED / Protected A / Protected B / Protected C / CONFIDENTIAL / SECRET / TOP SECRET) — replace the standard UK line in the header..arckit/references/citation-instructions.md. The Official Languages Act, An Act for the Substantive Equality of Canada's Official Languages (Bill C-13, 2023), the TBS Policy on Official Languages, and the Directive on the Implementation of the Official Languages (Communications with and Services to the Public) Regulations MUST appear in the Document Register with their primary URLs and verification dates.projects/<project-id>/<filename>.Official Languages Act (R.S.C., 1985, c. 31 (4th Supp.)), as amended by An Act for the Substantive Equality of Canada's Official Languages (S.C. 2023, c. 15). TBS Policy on Official Languages and the Directive on the Implementation of the Official Languages (Communications with and Services to the Public) Regulations. URL: <https://laws-lois.justice.gc.ca/eng/acts/O-3.01/>. Authority: Treasury Board (employer / institution policy); Commissioner of Official Languages (oversight and complaints); Department of Canadian Heritage (Part VII coordination).
After completing this command, consider running:
$arckit-ca-gc-digital-standards -- OLA service equivalence is a baseline expectation under the GC Digital Standards conformance scorecard.$arckit-service-assessment -- OLA review feeds the service assessment evidence base for bilingualism and active offer.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 3,445 | 58,079 | +1586% | 1 | 1 | 0% | 233 | 7,851 | +3270% | 0 | 0 | — |
case-02 | fail→fail | 16,598 | 6,598 | -60% | 1 | 1 | 0% | 2,118 | 1,921 | -9% | 0 | 0 | — |
case-03 | fail→fail | 19,362 | 5,451 | -72% | 1 | 1 | 0% | 3,152 | 1,758 | -44% | 0 | 0 | — |
case-04 | pass→fail | 30,193 | 32,260 | +7% | 1 | 1 | 0% | 6,201 | 7,731 | +25% | 0 | 0 | — |
case-05 | pass→fail | 23,422 | 8,664 | -63% | 1 | 1 | 0% | 4,140 | 2,513 | -39% | 0 | 0 | — |
case-06 | pass→fail | 25,378 | 6,724 | -74% | 1 | 1 | 0% | 4,429 | 2,019 | -54% | 0 | 0 | — |
case-07 | fail→pass | 10,297 | 17,688 | +72% | 1 | 1 | 0% | 1,782 | 4,432 | +149% | 0 | 0 | — |
case-08 | pass→pass | 11,272 | 4,077 | -64% | 1 | 1 | 0% | 1,803 | 2,240 | +24% | 0 | 0 | — |
case-13 | pass→fail | 12,769 | 4,347 | -66% | 1 | 1 | 0% | 2,140 | 1,885 | -12% | 0 | 0 | — |
case-09 | pass→fail | 15,577 | 5,875 | -62% | 1 | 1 | 0% | 1,922 | 1,890 | -2% | 0 | 0 | — |
case-10 | pass→fail | 11,790 | 6,825 | -42% | 1 | 1 | 0% | 1,837 | 2,035 | +11% | 0 | 0 | — |
case-11 | fail→fail | 13,417 | 5,746 | -57% | 1 | 1 | 0% | 2,068 | 1,813 | -12% | 0 | 0 | — |
case-12 | fail→pass | 16,844 | 14,391 | -15% | 1 | 1 | 0% | 2,965 | 3,800 | +28% | 0 | 0 | — |
case-19 | fail→pass | 17,365 | 13,763 | -21% | 1 | 1 | 0% | 2,442 | 3,657 | +50% | 0 | 0 | — |
case-14 | pass→pass | 7,469 | 15,596 | +109% | 1 | 1 | 0% | 1,256 | 3,770 | +200% | 0 | 0 | — |
case-15 | fail→fail | 11,615 | 6,543 | -44% | 1 | 1 | 0% | 1,841 | 2,023 | +10% | 0 | 0 | — |
case-16 | fail→pass | 19,012 | 25,295 | +33% | 1 | 1 | 0% | 2,846 | 4,369 | +54% | 0 | 0 | — |
case-17 | fail→fail | 10,304 | 7,363 | -29% | 1 | 1 | 0% | 1,587 | 1,962 | +24% | 0 | 0 | — |
case-18 | pass→pass | 13,317 | 12,183 | -9% | 1 | 1 | 0% | 2,131 | 3,453 | +62% | 0 | 0 | — |
case-20 | fail→pass | 8,392 | 2,718 | -68% | 1 | 1 | 0% | 1,313 | 1,799 | +37% | 0 | 0 | — |
case-21 | fail→pass | 15,163 | 4,355 | -71% | 1 | 1 | 0% | 2,131 | 2,272 | +7% | 0 | 0 | — |
case-22 | pass→pass | 14,206 | 13,080 | -8% | 1 | 1 | 0% | 2,327 | 3,465 | +49% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 11 counted toward the lift figure. The other 11 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of 0 percentage points is the difference between those two pass rates over the 11 comparable cases. 9 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.